An abstraction atop PyMongo for working with Django pythonically (or so we hope).
Switch branches/tags
Nothing to show
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.



An abstraction atop PyMongo for working with Django pythonically (or so we hope).


I'm developing with mongodb 1.4.3 and pymongo 1.6. Lesser versions will not work. Django is not necesarily required but is assumed here (1.2+)


Put djongobj on your path in your favored way.


You can address a Document with a db, collection, and unique id (optionally host and port). djongobj.models.Document acts like a dictionary. Each operation eagerly interacts with the db.

from djongobj.models import Document
d = Document('content', 'blog', 1)

A collection can be addressed with the db name and the collection name. A djongobj.models.Collection is a lightly-wrapped pymongo.collection.Collection.

from djongobj.models import Collection
c = Collection('content', 'blog')

With Django

A simple way to use the module is to add an instance of Mongo to your normal Django class:

from djongobj.models import Mongo

class Blog(models.Model):

    mongo = Mongo()

Using this descriptor, you can add in some arguments if you must:

Mongo(database='mydatabase', collection='mycollection', host='12.XX...', port=27017,
        document_class=MyDocument, collection_class=MyCollection)

MyDocument and MyCollection can be subclasses of djongobj.models.Document and djongobj.models.Collection respectively.

For the sake of these samples, we will assume that Mongo was instantiated without arguments.

Class attr - Collection

Now, with your class, the mongo attr will be a lightly-wrapped pymongo collection:

>>> Blog.mongo
<class 'djongobj.models.Collection'>, {'host': 'localhost', 'db': 'content', 'port': 27017, 'collection': 'blog'}
>>> Blog.mongo.all()
<pymongo.cursor.Cursor object at 0x9b5040c>

Do some basic queries:

>>> Blog.mongo.filter(_id=1)
<pymongo.cursor.Cursor object at 0x9b5338c>
>>> # get the query back as a queryset from the Django ORM:
>>> Blog.mongo.filter(_id=1, return_type='queryset')
[<Blog: this is the title>]
>>> len(Blog.mongo) # number of documents in the collection

Update some documents:

>>> Blog.mongo.update_one(1, {'foo': [12, 32]})
>>> Blog.mongo.upsert_one({'_id': 1, 'foo': [12, 32]}) # same as above in this case
>>> Blog.mongo.update({'foo': [12, 32]}, # the query params
...     {'$set': {'bar': 'baz'}}         # the update instructions
... )

Create some indices:

>>> from pymongo import ASCENDING, DESCENDING, GEO2D
>>> Blog.mongo.create_index(
... [ ('somekey', ASCENDING), ('other', DESCENDING), ('mypoint', GEO2D) ])

Advanced queries and map reduce are facilitated:


You can get at the pymongo objects underneath:

>>> Blog.mongo._collection
Collection(Database(Connection('localhost', 27017), u'content'), u'blog')
>>> Blog.mongo._db
Database(Connection('localhost', 27017), u'content')

Instance attr - Document

The document that is currently implemented here acts just like a dictionary that eagerly interacts with the database. On the surface, it looks like you are just interacting with regular dictionary but each operation incurs database traffic. This may or may not be desirable depending on the case. Input is welcome!

The pk field for SQL database and our collection are the same:

>>> b = Blog.objects.get(id=1) # get an object.
>>> b.mongo
{u'_id': 1, u'bar': u'baz'}
>>> b.mongo['foo'] = 'the next bar'
>>> del(b.mongo['bar'])